10050
J. Phys. Chem. A 2004, 108, 10050-10059
Vibrational and Electronic Spectroscopy of a Donor-Acceptor Substituted Distyrylbenzene and Its Covalent Dimers Weinan Leng,† Jason Grunden,‡ Glenn P. Bartholomew,§ Guillermo C. Bazan,§ and Anne Myers Kelley*,† School of Natural Sciences, UniVersity of California, Merced, P.O. Box 2039, Merced, California 95344, Department of Chemistry, UniVersity of California, Santa Barbara, Santa Barbara, California 93106-9510, and Department of Chemistry, Kansas State UniVersity, Manhattan, Kansas 66506 ReceiVed: June 22, 2004; In Final Form: September 1, 2004
Absorption spectra, resonance Raman spectra and depolarization ratios, fluorescence spectra and emission polarizations, and simulations of the resonance Raman excitation profiles and absorption spectra are reported for a donor-acceptor substituted distyrylbenzene (DADSB) and two covalent “dimers” formed by joining two DADSB chains at their center phenyl ring through a paracyclophane moiety. Semiempirical and densityfunctional theory calculations of the ground-state geometries and ZINDO calculations of the electronic excitations are also reported. Both the spectroscopic results and the electronic structure calculations agree that the electronic states of these “dimers” are not adequately described by excitonic coupling between the nominally degenerate electronic transitions localized on each DADSB moiety. The resonance Raman spectra of both dimers are essentially identical but show additional lines and intensity differences relative to the monomer. The excitation profiles of all three molecules exhibit interference effects between the Raman amplitudes for the first two strongly allowed electronic transitions. All three molecules exhibit Raman depolarization ratios of 1/3 throughout the lowest-energy absorption band, but the fluorescence of both dimers immobilized in polymer matrixes is considerably depolarized relative to that of the monomer. This suggests that the electronic excitation, initially delocalized over both conjugated chains of the dimers, may become localized on a single chain as geometric relaxation and solvent reorganization occur.
Introduction Many naturally occurring supramolecular structures, as well as man-made molecular materials, consist of noncovalent aggregates of conjugated molecules. In such aggregates, the intermolecular interactions are often weak enough to be considered as perturbations on the properties of the isolated molecules, yet strong enough that these perturbations are far from negligible. In many of these systems, such as the lightharvesting antenna pigments involved in photosynthesis,1,2 the thin films of conjugated polymers used in organic light-emitting diodes,3-6 and the donor-acceptor substituted poled polymers for second-order nonlinear optical applications,7-10 it is the optical properties of the aggregates that are of principal interest. The simplest models for the electronic spectroscopy of such aggregates assume that no electron transfer or exchange occurs between monomers, in which case the intermolecular interaction simply involves the charge distributions on the individual monomers. For strongly allowed electronic transitions, the dominant term is the coupling between transition dipole moments on the constituent monomers, which splits the initially degenerate electronic excitations on the monomers into a band of transitions whose energies and oscillator strengths depend on the positions and relative orientations of the interacting molecules.11-13 There are some systems, such as the special pair in photosynthesis, in which the aggregates have a well-defined geometry. * To whom correspondence should be addressed. † University of California, Merced. ‡ Kansas State University. § University of California, Santa Barbara.
Many aggregates, however, contain a large number of randomly or semi-randomly positioned monomers encompassing a wide variety of different local interaction geometries. This heterogeneity presents a challenge for understanding the details of the intermolecular interactions or for predicting the optical properties of the aggregate. Many workers have therefore focused on simpler systems in which a small number of chromophores, often only two, are arranged in a well-defined geometry. Although noncovalent interactions sometimes provide dimers or even larger aggregates that appear to have a welldefined composition and geometry,14-16 connecting the monomers through a covalent but electronically nonconjugated linkage is a more general and more certain way to force a single stoichiometry and spatial arrangement. One convenient way to generate covalent dimers of monomers containing phenyl groups is to attach them through a paracyclophane linkage. The Bazan group, in particular, has synthesized and examined a number of such dimers based on stilbene, distyrylbenzene, and longer members of this series.17-21 The original covalent dimers of nonpolar monomers were intended as model systems for the intermolecular interactions in para(phenylene vinylene) and its substituted analogues used as organic LEDs. More recently, these studies have been extended to electron donor-acceptor substituted stilbenes and distyrylbenzenes.22-25 These are interesting systems for addressing questions of through-space versus through-bond charge transfer and can be considered as model systems for the “push-pull” conjugated molecules useful as organic second-order nonlinear optical materials. This paper describes spectroscopic and computational studies on a donor-acceptor substituted distyrylbenzene monomer and
10.1021/jp047280f CCC: $27.50 © 2004 American Chemical Society Published on Web 10/27/2004
Donor-Acceptor Substituted Distyrylbenzene
J. Phys. Chem. A, Vol. 108, No. 46, 2004 10051
Figure 1. Monomer and two dimers studied in this work.
two of its paracyclophane-linked dimers that differ only in the relative orientation of the two monomers (Figure 1). The synthesis of these molecules and their absorption and fluorescence spectra in solution have already been reported,24 as have their dipole moments and first hyperpolarizabilities.25 Here we report and analyze their resonance Raman spectra and excitation profiles in solution and their fluorescence spectra and anisotropies in polymer films and present electronic structure calculations on both the ground-state geometry and the electronic excitations. Experimental Methods The synthesis and characterization of DADSB, 5, and 6 have been described elsewhere.24,25 Absorption spectra were measured on a Hitachi U-3010 UV/ vis spectrophotometer in 1 mm path length cells. Molar absorptivities were determined for DADSB, 5, and 6 in CH2Cl2 by measuring spectra of carefully weighed samples. Absorptivities in CHCl3 were determined by recording spectra of equal dilutions of CH2Cl2 stock solutions into CH2Cl2 and CHCl3. These absorptivities were then used to determine the concentrations of the samples used for Raman spectroscopy. Resonance Raman spectra of DADSB, 5, and 6 were obtained in both CH2Cl2 and CHCl3 solution at concentrations ranging from 0.03 to 0.6 mM. Two different excitation, sampling, and detection systems were employed. Spectra in CH2Cl2 were obtained on a 0.6-m Spex 1877E triple spectrograph utilizing reflective collection optics, a 1200 g/mm grating blazed at 500 nm in the spectrograph stage, and a liquid nitrogen cooled CCD detector. Samples of ∼3 mL volume were contained in a spinning cell, illuminated with 1-20 mW of laser power focused with a 10 cm f.l. lens, and the Raman scattering collected in a ∼135° backscattering geometry. Excitation wavelengths of 458, 488, and 514 nm were provided by a Lexel argon-ion laser, and excitation at 543 nm was obtained from a green He-Ne laser. Spectra in CHCl3 were obtained on a 0.64-m Jobin-Yvon T64000 triple spectrograph with microprobe sampling (10× objective), either a 1200 g/mm grating blazed at 750 nm or an 1800 g/mm grating blazed at 500 nm in the spectrograph stage, and a UV coated, back-illuminated, liquid-nitrogen-cooled CCD detector. Samples of ∼10 mL volume were circulated with a peristaltic pump through a 1 mm path length liquid flow cell, and Raman scattering was collected in a confocal 180° backscattering geometry. Excitation was provided by a Coherent Innova 90C argon-ion laser (363 nm and 458-514.5 nm), an argon ion-pumped Coherent 599 dye laser (563 nm), and a Spectra-Physics Tsunami Ti:sapphire laser producing 1-2 ps
pulses at 82 MHz, frequency doubled to 400-450 nm. Laser powers at the sample were 1-10 mW for argon or dye laser excitation and no greater than 1 mW for Ti:sapphire excitation. In both instruments, the scattering was collected at 90° to the incident laser polarization direction and the scattered polarization was randomized by passage through a polarization scrambler placed before the spectrometer entrance slit. The intensities measured therefore correspond to the differential Raman cross section, (dσ/dΩ)|+⊥. For the Raman polarization measurements, the laser polarization was purified by passage through a GlanTaylor prism, and a rotatable film polarizer was placed in the scattered beam path before the polarization scrambler. Spectral resolution was 4-7 cm-1 depending on wavelength. Spectra were calibrated in Raman shift by reference to Raman lines of the solvent. Intensities were corrected for the spectral response of the spectrograph and detector by collecting spectra of a tungsten-halogen standard lamp as described elsewhere.26 Intensities were corrected for reabsorption of the scattered light by using the algorithm derived in ref 27 for the experiments in CH2Cl2, where the samples were optically thick, and using a variation of the algorithm presented in ref 28 for the experiments in CHCl3, which were performed at lower concentrations. Integrated peak areas were determined by fitting regions of the spectrum to sums of mixed Gaussian-Lorentzian peaks (Grams32) after subtraction of a low-order polynomial to remove underlying fluorescence. The absolute differential resonance Raman cross sections were determined by measuring the integrated areas of the chromophore Raman bands relative to that of the 702 cm-1 line of the CH2Cl2 solvent29 or the 667 cm-1 line of the CHCl3 solvent.30 The A-term fitting parameters for CHCl3, which refer to total Raman cross section, were converted to differential cross section using26
dσ (dΩ )
) |+⊥
(8π3 )(11++2FF )σ
T
(1)
with F ) 0.02 for the 667 cm-1 line of CHCl3. Resonance Raman depolarization ratios were measured for 5 and 6 in CH2Cl2 at 424, 458, and 514.5 nm excitation and in CHCl3 at 563 nm. The incident laser polarization was purified by passage through a Glan-Taylor prism and the Raman scattering was detected through a sheet polarizer (Oriel) placed before the polarization scrambler. The detected polarization was alternated several times and the spectra at each polarization were summed. In the T64000 system, where the scattered light passes through a beam splitter at 45° before its polarization is scrambled, the intensities were corrected for the polarization dependence of the collection efficiency by collecting “parallel”
10052 J. Phys. Chem. A, Vol. 108, No. 46, 2004 and “perpendicular” spectra of an unpolarized source (a tungsten-halogen lamp scattered off a BaSO4 plate). For the fluorescence measurements, samples in polystyrene (PS; Aldrich, MW 760-770) and poly(methyl methacrylate) (PMMA; Aldrich, MW 120,000) films were prepared as described previously31 at chromophore concentrations of 10-5 to 10-6 by mass. Fluorescence spectra were obtained on a homebuilt fluorometer. The excitation source was a 150 W Xe-Hg lamp followed by an Oriel 0.125-m double monochromator. The detection system was a 0.65-m ISA spectrograph and a liquid nitrogen cooled Princeton Instruments CCD. The films were optically thin, and excitation and detection were nearly collinear. For anisotropy measurements, the excitation light was linearly polarized with one Polaroid sheet and parallel or perpendicularly polarized emission was selected with a second sheet, and a polarization scrambler was placed before the spectrograph entrance slit. All experiments were performed at ambient temperature. Fluorescence spectra have not been corrected for instrument response. Computational Methods Energy minimizations, normal mode calculations, and electronic structure calculations were carried out using the Gaussian 98 suite of programs32 running under Windows. The hexyl groups were replaced by methyl groups. Ground-state geometries were calculated for all three molecules with the AM1 and PM3 semiempirical Hamiltonians as well as density functional theory with the B3LYP hybrid density functional and two basis sets, STO-3G and 6-31G. Ground-state vibrational normal mode calculations were carried out for DADSB only using density functional theory with the 6-311G** basis set. Molden 3.6 was used to animate and visualize the normal modes.33 Calculations of the electronic spectra were carried out using the ZINDO semiempirical method. The excited states were calculated using configuration interaction among all singly excited configurations formed from all 134 molecular orbitals for DADSB and from the 31 highest occupied and 50 lowest virtual orbitals for dimers 5 and 6. Only the six (for DADSB) or 10 (for 5 and 6) lowest electronic transitions were calculated. The absorption and resonance Raman spectra were simulated via the time-domain wave packet method modified to consider contributions to the resonance enhancement from two different electronic states.34,35 Each Raman-active mode was treated as a harmonic oscillator characterized by its frequency and a displacement ∆, in dimensionless normal coordinates, between the potential minima in the ground state and a given excited state. Changes in vibrational frequency upon excitation, mixing of the normal modes in the excited state (Duschinsky rotation), and coordinate dependence of the electronic transition moment were not considered. The electronic line width was partitioned into inhomogeneous broadening, modeled as a static Gaussian distribution of electronic zero-zero energies, and homogeneous broadening, described by coupling of the electronic transition to an overdamped Brownian oscillator representing the solvent degrees of freedom. Complete correlation between the shifts of both electronic transitions within the inhomogeneous distribution was assumed36 (but see also ref 34). The parameters were adjusted to obtain the best simultaneous fit to the absorption spectrum and the absolute resonance Raman cross sections at all measured wavelengths. The second absorption band (near 370 nm) may be composed of more than one electronic transition, and the longest-wavelength band probably also contains more than one transition, at least in the dimers (vide infra). However, the wavelength-dependent Raman intensities give direct evidence for only two contributing transitions.
Leng et al.
Figure 2. Absorption spectra in CH2Cl2 (units on y axis) and fluorescence spectra in PMMA films (arbitrarily scaled) for DADSB (solid black), dimer 5 (solid gray), and dimer 6 (dashed). Fluorescence excitation wavelengths are 450 nm for DADSB and 475 nm for the dimers.
Results Absorption Spectra. Figure 2 shows the linear absorption spectra of DADSB, 5, and 6 in CH2Cl2. The monomer has a strong transition at about 22 500 cm-1 and a broader band, which may encompass more than one transition, at 26 000 to 30 000 cm-1. The two dimers, 5 and 6, have very similar spectra with the low-energy band shifted to the red by 1000-2000 cm-1 and some changes in the shape of the higher-energy band(s). Notice that the integrated molar absorptivity of the dimers is only slightly greater than that of the monomer even though the dimers contain two DADSB chromophores. Fluorescence Spectra and Anisotropies. The fluorescence spectra in fluid solution were reported in ref 24. To gain further insight into the nature of the absorbing and emitting states, we measured fluorescence spectra and anisotropies of the chromophores immobilized in PMMA and PS matrixes. Here the chromophores are randomly oriented but have little ability to rotate between the light absorption and emission steps. The fluorescence spectra in PMMA, excited near the absorption maxima, are shown in Figure 2. All three molecules have very similar spectra, with the dimer spectra slightly red-shifted relative to that of the monomer. The fluorescence spectra in hexanes, which show some vibronic structure, are also quite similar for all three molecules.24 The emission maxima in PMMA show a significant dependence on excitation wavelength as shown in Figure 3. For the monomer, excitation at shorter wavelengths leads to emission at shorter wavelengths with excitation in the 400-500 nm range. This behavior is consistent with the existence of an inhomogeneous distribution of absorbers that differ in the local polarity or refractive index of the medium and/or the planarity of the chromophore.37 Excitation at different wavelengths photoselects different subsets of the inhomogeneous population which do not interconvert during the fluorescence lifetime, so redder excitation produces redder emission. At even shorter excitation wavelengths (375-350 nm), the higher-energy band starts to contribute significantly to the absorption and the contribution of redder-emitting molecules increases again. The dimers show the same trend for excitation within the low-energy absorption band at 425-550 nm, but at shorter excitation wavelengths (400-350 nm), where the second transition should start to contribute significantly, the emission further blue-shifts.
Donor-Acceptor Substituted Distyrylbenzene
Figure 3. Dependence of fluorescence maximum on excitation wavelength for DADSB and dimers 5 and 6 in PMMA films at room temperature.
Figure 4. Fluorescence anisotropy as a function of excitation wavelength for DADSB and dimers 5 and 6 in PMMA at room temperature.
Figure 4 plots the fluorescence anisotropy, r ) (I| - I⊥)/ (I| + 2I⊥), as a function of excitation wavelength in PMMA. The anisotropies at a given excitation wavelength are only weakly dependent on emission wavelength and are nearly constant over the region where there is significant emission. For the monomer, excitation within the main absorption band (400-500 nm) results in a fluorescence anisotropy only slightly below the value of r ) 0.4 expected for a single electronic transition of a randomly oriented sample unable to rotate during the fluorescence lifetime. The slight deviation from 0.4 may indicate a small degree of rotational freedom within the polymer matrix. Both dimers show considerably lower anisotropies even when excited within the main absorption band (450-550 nm). Similar anisotropy data were obtained in PS films, although in general the anisotropies were slightly lower in PS. Resonance Raman Spectra and Depolarization Ratios. Figure 5 displays the resonance Raman spectra of all three molecules in CH2Cl2 at an excitation wavelength of 458 nm. The dimer spectra are virtually identical to each other and differ from the monomer spectrum mainly in the “fingerprint” region around 1150-1300 cm-1. The frequencies and relative intensities are nearly independent of excitation wavelength within the main absorption band (above ∼430 nm for the monomer and above ∼450 nm for the dimers). Figure 6 shows the Raman polarization data of the dimers with excitation at 458 nm. For all Raman lines, the depolarization ratio F ) I⊥/I|| is nearly identical to 0.33, the value expected for resonance with a single allowed electronic transition. Essentially identical results were obtained using excitation at 514.5 nm and on the red edge of the absorption at 563 nm, but a slightly higher value of F ) 0.4
J. Phys. Chem. A, Vol. 108, No. 46, 2004 10053
Figure 5. Resonance Raman spectra of DADSB and dimers 5 and 6 in CH2Cl2 solution at 458 nm excitation. Intensity scale is arbitrary. Fluorescence backgrounds have been subtracted and dimer spectra are offset vertically for clarity.
Figure 6. Polarized resonance Raman spectra of dimers 5 and 6 in CH2Cl2 at 458 nm excitation. Fluorescence backgrounds have been subtracted from all spectra and the perpendicular spectra multiplied by a factor of 3. Relative scaling and vertical offset of the 5 and 6 spectra is arbitrary. The parallel and perpendicular*3 spectra are nearly indistinguishable.
Figure 7. Resonance Raman spectra of dimer 5 in CHCl3 at three excitation wavelengths. Vertical scales and vertical offsets are arbitrary.
was obtained for both 5 and 6 when excited at 424 nm, where the higher-energy band is starting to contribute to the absorption. Figure 7 compares the Raman spectra of dimer 5 obtained with excitation on the red side of the first absorption band, on the blue side of that band, and within the second absorption band. The spectral patterns are nearly the same throughout the main absorption band, but clearly become very different when exciting into the higher-energy band. This indicates that the geometry changes induced by excitation into the higher-energy and lower-energy bands are quite different, and localized on different parts of the molecule.
10054 J. Phys. Chem. A, Vol. 108, No. 46, 2004
Leng et al.
TABLE 1: Calculated Geometries for DADSB and Dimers 5 and 6 from Different Methods, and Calculated (ZINDO) Electronic Transition Wavelengths, Oscillator Strengths, and Polarizations at Each Geometry methoda
AM1 stateb
DADSB geomd
dimer 6
geomd
dimer 5
geomd
nm
PM3 f, pol.c
3 409 1.95, x 4 345 0.09, xy spiral twisted 0°/45°/90°, rCedCe)1.34, rNO)1.20, rCe-Cphen)1.45, µ)3.8 D 4 7 8 9
438 0.89, x 406 0.19, y 390 0.92, x 381 0.20, x rings twisted ∼45°, slightly bent; rCedCe)1.34, rNO)1.20, rCe-Cphen)1.45, µ)3.3 D 4 7 8
437 410 389
0.95, x 0.22, y 1.16, x
NO2 rings planar, Nhex2 rings twisted ∼30°; rCedCe)1.34, rNO)1.20, rCe-Cphen)1.45, µ)6.1 D
stateb 3 4
4 5 8
nm
B3LYP/STO-3G f, pol.c
403 1.83, x 342 0.13, x planar, rCedCe)1.34, rNO)1.23, rCe-Cphen)1.45 443 423 401
1.12, x 0.15, y 0.96, x
stateb
371 0.36, x 337 0.78, y 331 0.15, y 328 0.23, z near planar, bent ∼120°; rCedCe)1.34, rNO)1.22, rCe-Cphen)1.46
B3LYP/6-31G f, pol.c
stateb
nm
f, pol.c
3 4
457 1.14, x 368 0.88, x planar, rCedCe)1.36, rNO)1.32, rCe-Cphen)1.49, µ)7.7 D
444 1.68, x 366 0.40, x planar, rCedCe)1.36, rNO)1.27, rCe-Cphen)1.46, µ)13.1 D
5 6 7 10
511 0.12, y 490 1.45, x 463 0.15, y 391 0.39, x near planar, slightly bent; rCedCe)1.36, rNO)1.32, rCe-Cphen)1.49, µ)7.1 D
3 486 0.16, y 6 459 2.11, x 7 439 0.16, y 10 377 0.31, x NO2 rings twisted 12°; rCedCe)1.36, rNO)1.27, rCe-Cphen)1.46, µ)11.8 D
one unit planar, other NO2 ring twisted ∼45°; rCedCe)1.34, rNO)1.22, rCe-Cphen)1.45-1.46 7 8 9 10
nm
5 6 7 10
515 0.13, y 479 1.60, x 463 0.13, y 387 0.42, x near planar, slightly bent; rCedCe)1.36, rNO)1.32, rCe-Cphen)1.48-1.49, µ)13.7 D
1 2
3 6 7 10
494 0.19, y 456 2.39, x 441 0.16, y 366 0.23, x NO2 and Nhex2 rings twisted
